prop.pass = proportion of students who passed the exam
num = number of students who sat the exam
urban = if school is in urban area (1 = urban, 0 = rural)
%dis = % pupils at school defined as "disadvantaged" (by parental income). Students either are or aren't disadvantaged.
Is the following regression suitable to determine if there is a difference in the rural and urban exam pass rate, controlling for %dis?
glm(prop.pass ~ urban + %dis, weights = num, family = quasibinomial, data = exam_df)
Secondly, I get the output:
Coefficients:
(Intercept) urban %dis
3.5123 -0.1244 -0.8605
Degrees of Freedom: 3156 Total (i.e. Null); 3154 Residual
(73 observations deleted due to missingness)
Null Deviance: 55930
Residual Deviance: 27650 AIC: NA
Do I interpret this as being from an urban rather than rural area is associated with an decrease in the log odds of passing the exam of 0.1244?
